So once you have, say, 30 or more (depending on whom you read and which type of chart you’re using) observations with no special cause signals, then you can calculate the capability indices using the sigma values from your chart. Cp is calculated by subtracting the lower specification limit from the upper specification limit, and dividing the result by six sigma units (basically, the distance represented by your control limits).

The “centered” capability index, or Cpk, is the smaller of two values, Cpu and Cpl.

Cpu is given by subtracting the process average from the upper specification limit, and dividing by 3 sigma.

Cpl is given by subtracting the lower specification limit from the process average, and dividing by 3 sigma.

You should NEVER assess capability in the absence of stability. ]]>

Sorry for the delay in responses. There’s no need for a table calculation . If your viz has hierarchy like Category, Subcategory, Product and there’s a Subcategory subtotal then MIN([Product]) = MAX([Product]) will detect the subtotals. Of course this shares the drawback of similar total & subtotal detection methods in that you’ll need to decide on what behavior to use if there’s only 1 product in a given subcategory.

]]>FIXED LOD expressions are computed before regular dimension filters are applied. In order to have the filters affect the results of the FIXED LOD expressions you’ll need to add the filters to context.See https://onlinehelp.tableau.com/current/pro/desktop/en-us/filtering_context.html for more details on context filters.

]]>I am facing the same issue where i need to calculate Total no of orders.

it should not change when use region as Filter.

Getting error with this

{FIXED : COUNT([Order ID])} ]]>

Thanks for this helpful post on using Level of Detail expressions to filter out top and bottom percentiles!

However, the calculation doesn’t work when we add additional fields to the Filters tab.

For example, we want to pull Order Date and City onto Filters and update our view to filter the top/bottom percentile for this new subset, but the percentile does not get updated by the filters and continues to refer to the total set of data.

I think this might be because the calculation has Percentile FIXED

{FIXED : PERCENTILE({FIXED [Customer Name] : SUM([Sales])}, .95)}

which is necessary, but introduces issues when you want to apply filters.

Would you be able to give advice on how to still fix the Percentile, but include certain desired filters?

I’ve tried using {INCLUDE [Order Date], [City]: } to various pieces of the calculation, but I haven’t gotten it to work yet.

Thanks for your help.

Thanks for this great info. Do you have an example using the WINDOW_MIN(MIN([Subcategory])) method above that identified subtotals rows & grand total rows? In my case I have a category & subcategory & I need to conditionally format the subtotal rows only.

]]>